GOOPAX 3.1.0 brings new powerful features that are especially useful for handling large, complex data structures, and for sharing data between multiple GPUs in parallel:

  • pointer arithmetics, where pointers can point to intrinsic types or to structs.
  • unified memory / shared virtual memory (sharing an address space across multiple devices)
  • fast inter-gpu communication

As always, everything is perfectly and elegantly integrated into the C++ language.

With GOOPAX version 3.0.1, the Windows version is also available, versions 3.0.2 and 3.0.3 brought stability fixes.

We are pleased to announce that GOOPAX 3.0 is available! Linux and Mac versions can be downloaded now, Windows versions will follow soon.

The new version offers many new features:

  • Multi-GPU support
  • CUDA backend and CUDA interoperability
  • Greatly improved support for images, image arrays and image samplers
  • fast data exchange between neighboring threads

… and many other features. We hope you like it.